Senkomago_unc_0153D_14076.pdf

91 

Full text

(1)

THEROLEOFHUMANPAPILLOMAVIRUS(HPV)VIRALLOADINPENILEHPV INFECTIONANDCLEARANCEAMONGYOUNGKENYANMEN

Virginia Senkomago

A dissertation submitted to the faculty of the University of North Carolina at Chapel Hill in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the

Department of Epidemiology.

Chapel Hill 2013

Approved by: Jennifer S. Smith Frieda Behets

(2)

© 2013

(3)

ABSTRACT

VIRGINIA SENKOMAGO: The role of human papillomavirus (HPV) viral load in penile HPV infection and clearance among young Kenyan men

(Under the direction of Jennifer S. Smith)

Persistent infections with human papillomavirus (HPV) types 16 and 18 are causes of cervical cancer in women and various penile squamous cell carcinomas (SCC) in men. Male circumcision has been found to be protective against penile HPV, but the association between circumcision and HPV viral load remains unclear.

Additionally, the role of HPV viral load in HPV persistence and subsequent development of penile SCC is unknown.

An HPV-ancillary study, nested within a randomized controlled trial (RCT) of male circumcision, was conducted in Kisumu, Kenya. Eligible participants were HIV seronegative, uncircumcised and aged 18-24. Penile swabs were collected from glans and shaft sites every 6 months for 24 months. GP5+/6+ PCR was used to identify HPV DNA types. HPV viral load was measured with LightCyler real-time PCR and classified as high (>250copies/scrape) or low (250copies/scrape).

(4)

baseline, risk of persistence to 6 months was lower in the circumcision arm [0.20(0.09-0.34)] than control arm [0.55(0.39-0.68)] for HPV16 and HPV18 [0.17(0.05-0.34) and 0.50(0.25-0.71), respectively].

In uncircumcised men, the hazard of HPV16 clearance at 6 months after first HPV16 detection was found to be lower for high versus low viral load incident infections in the glans [adjusted hazard ratio (aHR) =0.58 (95%confidence interval, 0.36-0.93)]. HPV16 and HPV18 clearance in the shaft was comparable for high and low viral load infections.

Male circumcision reduces the acquisition and possibly enhances the clearance of high viral load HPV16 or HPV18 infections in the glans, and thus could potentially reduce HPV transmission to women. The reduced rate of high versus low viral load HPV16 clearance in uncircumcised men could be associated with increased

(5)

ACKNOWLEDGEMENTS

I would first like to thank my advisor and dissertation committee chair, Jennifer Smith, for introducing me to HPV research and for her guidance, patience,

encouragement and commitment through every step of this project. Her hard work and dedication to HPV research and cervical cancer prevention is truly inspiring. I would also like to sincerely thank Michael Hudgens for his statistical expertise, dedication to this project and frequent accessibility when I needed advice. I am very thankful to Charlie Poole for his guidance in epidemiology methods and for his advice throughout this project. I am also very grateful to Freida Behets and Steve Meshnick for their guidance and helpful comments during my entire dissertation process.

(6)

I am immensely thankful to Michael Schwarz, Andrew Waters and the Tellus Educational Foundation for their encouragement and financial support with the Tellus

Foundation Scholarship throughout my doctoral education. I am also very thankful to Louise and Derek Winstanly for the Winstanly scholarship awarded to me in my first year in the doctoral program. I also want to acknowledge the Armand Hammer United World College and Berea College scholarships that supported my previous education

and inspired me to pursue a career in global health improvement.

I also want to thank Vangie Foshee, Wendee Wechsberg and Prema Menezes who provided me with research opportunities during my doctoral education and enabled me to further build my research and management skills. I also want to thank Nancy Colvin, Carmen Woody and Bobbi Wallace for their guidance.

Lastly, I want to thank my family, who even from so far away have been my biggest cheerleaders and provided unwavering love and encouragement throughout my education. I am also truly grateful to all my friends at UNC and beyond for the

(7)

TABLE OF CONTENTS

LIST OF TABLES ...ix

LIST OF FIGURES ... x

LIST OF ABBREVIATIONS………....xi

CHAPTER 1. STATEMENT OF SPECIFIC AIMS ... 1

CHAPTER 2. BACKGROUND AND SIGNIFICANCE... 5

Significance of HPV infections in males ... 5

The role of Male HPV infections in female HPV and cervical cancer ... 7

Male Circumcision and HPV prevalence, incidence, clearance and transmission ... 8

The Role of HPV Viral Load in the Natural History of HPV infections ... 9

Biological Plausibility for Proposed Hypotheses ... 11

CHAPTER 3. METHODS ... 14

RCT of Male Circumcision for HIV Prevention: Bailey et al, 2007 ... 14

Design of HPV Nested Study ... 16

(8)

CHAPTER 4. THE ASSOCIATION BETWEEN MALE CIRCUMCISION

AND PENILE HPV VIRAL LOAD IN AN RCT IN KISUMU, KENYA ... 26

Overview ... 26

Introduction ... 27

Methods ... 28

Results ... 33

Discussion ... 37

Tables and Figures……….41

CHAPTER 5. ASSOCIATION BETWEEN PENILE HPV VIRAL LOAD AND HPV CLEARANCE IN UNCIRCUMCISED KENYAN MEN ... 50

Overview ... 50

Introduction ... 51

Methods ... 52

Results ... 58

Discussion ... 60

Tables and Figures………....65

CHAPTER 6. CONCLUSION ... 72

(9)

LIST OF TABLES

Table Page

Table 4.1: Baseline Characteristics by Intervention Treatment Arm………..45

Table 4.2: Estimated effect of circumcision on incident high viral load

HPV16 and HPV18 infections over 24 months in Kenyan Men………...47

Table 4.3: Association between circumcision and prevalence of high

versus low viral load at detection of incident infections………48

Table 4.4: Clearance of baseline prevalent infections by circumcision

status and HPV viral load in Kenyan Men………...……49

Table 5.1: Participant characteristics, stratified by baseline HPV viral load

for HPV types 16 and 18 in men in Kisumu, Kenya………..…68

Table 5.2: The association of penile HPV viral load and HPV clearance in

(10)

LIST OF FIGURES

Figure Page

Figure 2.1: Natural History of HPV Infections showing HPV viral

replication process………..13

Figure 3.1: Directed Acyclic Graph (DAG) to Assess Confounders for

Potential Confounders……….……….25

Figure 4.1: Study flowchart of HPV samples included in analyses of

male circumcision and HPV viral load………..………..41

Figure 4.2: Estimated Cumulative Incidence of High Viral Load infections

over 24 months in Kenyan Men………..42

Figure 4.3: Estimated Clearance of Prevalent High Viral Load HPV

Infections by circumcision status in Kenyan Men………....43

Figure 4.4: Estimated Clearance of Low Viral Load Prevalent HPV

Infections by circumcision status in Kenyan Men………...…..44

Figure 5.1: Study flowchart of samples included in analyses of HPV

Viral Load and HPV clearance………65

Figure 5.2: Estimated clearance of prevalent HPV infections by HPV Viral

Load in uncircumcised Kenyan Men………. 66

Figure 5.3: Estimated clearance of incident HPV infections by HPV Viral

(11)

LIST OF ABBREVIATIONS

CI Confidence interval

CIN Cervical intraepithelial neoplasia ELISA Enzyme-linked immunosorbent assay DAG Directed Acyclic Graph

HIV Human immunodeficiency virus HPV Human papillomavirus

HSV-2 Herpes simplex virus type 2

HR Hazard Ratio

ICC Invasive cervical cancer PCR Polymerase chain reaction PR Prevalence Ratio

RCT Randomized controlled trial RLB Reverse line blot

SCC Squamous Cell Carcinoma STI Sexually transmitted infection

(12)

CHAPTER 1

STATEMENT OF SPECIFIC AIMS

Persistent infections with human papillomavirus (HPV) high-risk types 16 and 18 are causes of approximately 70% of cervical cancer cases (1). Higher HPV16 or HPV18 viral loads in men are believed to be associated with increased HPV transmission female partners compared to lower viral loads (2). Male circumcision has been found to be protective against HIV and penile HPV incidence,however, the effect of circumcision on HPV viral load of incident and prevalent HPV infections remains unclear (3-8).

In men, persistence of HPV16 and HPV18 infections is believed to cause penile squamous cell carcinomas (SCC), particularly basaloid and warty SCC types (9, 10). Higher HPV16 or HPV18 viral loads have been associated with decreased HPV

clearance and increased progress to high grade cervical intraepithelial neoplasia (CIN) in women (11-17). However, the association between HPV viral load and HPV

clearance has not yet been explored in penile HPV infections in men.

(13)

real-time PCR assay; viral load was classified as high or low (>250 versus 250

copies/scrape) (21). HPV Incidence was defined as the first HPV16 or HPV18-positive result in the glans or shaft. HPV clearance was defined as an HPV-negative result in participants who were positive for that given HPV type in the same anatomical site (glans or shaft) at their last visit.

With the data collected from this study, the following research aims were addressed:

Aim 1: To examine the association between male circumcision and HPV16 or HPV18 viral load in Kenyan men aged 18-24.

1.1. To examine the association between circumcision and the incidence of high viral load infections for HPV types 16 and 18 in the glans or shaft.

1.2. To examine the association between circumcision and the prevalence of high versus low HPV viral load infections among men with incident HPV 16 or HPV18 infections in the glans or shaft.

1.3. To examine the association between circumcision and clearance of prevalent high and low viral load infections for HPV types 16 and 18 in the glans or shaft

Hypotheses:

(14)

Rationale: Circumcision and lower HPV viral load in men are believed to be associated with reduced HPV transmission to women (2). The effect of circumcision on HPV viral load, however, remains unclear. Only one study, an RCT of male circumcision in

Uganda, has examined the association between male circumcision and penile HPV viral load (22).This study did not have longitudinal data on HPV viral load and was therefore unable to examine the association between circumcision and the incidence of high viral load HPV infections, or the association between circumcision and the clearance of high and low viral load prevalent HPV infections. Higher HPV viral loads in men are possibly associated with greater HPV transmission to their female partners (2); thus,

circumcision may reduce the transmission of HPV infections to women by reducing incidence and enhancing clearance of high viral load HPV infections in their male partners.

Aim 2: To examine the association between viral load at HPV detection and HPV clearance in uncircumcised Kenyan men aged 18-24.

2.1. To examine the association between HPV viral load (high versus low) at detection of prevalent HPV 16 and 18 infections and HPV clearance in the glans or shaft

2.2. To examine the association between HPV viral load (high versus low) at detection of incident HPV 16 and 18 infections and HPV clearance in the glans or shaft

Hypothesis

(15)

Rationale: High viral load of high-risk HPV types is suggested to be associated with increased HPV persistence in women; however the association between HPV viral load and HPV clearance has not been examined among penile HPV infections in men (11-17). Higher HPV16 or HPV18 viral loads could possibly be associated with lower HPV clearance and subsequently increased progression to penile SCC.

(16)

CHAPTER 2

BACKGROUND AND SIGNIFICANCE

Significance of HPV infections in males

HPV is one of the most common sexually transmitted infections (STIs) worldwide, the overall global prevalence of HPV in men ranges from 1% to 84% among low-risk and from 2% to 93% among high-risk male populations (23). In Africa, HPV prevalence in males has been found to be as high as 58% (23). The overall HPV prevalence at baseline in the men who participated in this study was 51%; the type-specific prevalences for HPV 16 and HPV 18 were 9.8% and 4.3% respectively (24). HPV infection in men is associated with penile, oral and anal cancers as well as HPV transmission to females and cervical cancer incidence (25). HPV infection in men has also been shown to be associated with an increased risk of HIV acquisition; data from men in this study found the risk of HIV acquisition in HPV-positive men to be 1.8 times that of HPV-negative men (26)

Determinants of HPV infection in males

(17)

lifetime sexual partners and no condom use. The presence of other sexually transmitted infections, particularly N. gonorrhea, C. trachomatis and herpes virus simplex virus type 2 (HSV-2) has also been associated with a higher prevalence of HPV in males (26). Personal characteristics of males including less than a secondary school education and less than daily bathing have also been associated with a greater HPV prevalence (24, 27, 28). Longitudinal studies examining factors associated with acquisition of HPV in males have found a greater number of lifetime sexual partners and infection with multiple HPV types to be associated with a greater risk of HPV acquisition (29, 30).

Determinants of HPV clearance in males

The average duration of clearance for overall HPV infections in males is reported to range from 5.9 - 7.52 months (31). Infection with HIV and multiple HPV types was found to be associated with decreased HPV clearance (29, 32). The results on the association of age with HPV clearance remain unclear; HPV clearance has to be found to increase in older men in one study (30), but was found to have no effect on clearance in another study (31).

HPV infection and Development of Penile Carcinomas

Persistence of HPV16 and HPV18 infections is a risk factor for the development of penile squamous cell carcinomas (SCC), particularly basaloid and warty SCC types (9, 10). Approximately half (45% - 47.9%) of penile carcinomas have detectable HPV DNA and the most common HPV types include HPV16 (30.8%) and HPV18 (6.6%) (33). Although invasive penile SCC is rare in comparison to other cancers, incidence in

(18)

over 5 times higher than in developed countries and is associated with high morbidity and mortality (33, 34).

The role of Male HPV infections in female HPV and cervical cancer

Even before the association between HPV and cervical cancer was understood, male sexual behavior was found to be associated to cervical cancer in female partners. Earlier studies found an association between a greater number of male sexual partners and cervical cancer even after adjustment for age of first intercourse and number of female sexual partners (35). This association generated the hypothesis of an infectious cause of cervical cancer.

(19)

Male Circumcision and HPV prevalence, incidence, clearance and transmission

The effect of male circumcision on HPV infections has been investigated by several studies. Two global meta-analyses including results from 2 RCTs and several longitudinal studies have examined the association between male circumcision and HPV prevalence, incidence and clearance (37, 38). Circumcision was found to greatly reduce the prevalence of overall penile HPV in both meta-analyses [Odds Ratio (OR) = 0.57 (0.45 -0.71) and OR= 0.57(0.42-0.77) respectively], particularly in the glans/sulcus [OR =0.47 (0.37-0.60)].

The two meta-analyses found different results on the association between male circumcision and acquisition of HPV infections; one meta-analysis found that

circumcision decreases HPV incidence [Rate Ratio (RR) = 0.75 (0.57-0.99)] and the other found no effect of circumcision on HPV incidence [OR = 1.01 (0.66-1.53)] (37, 38). However, results from RCTs in Rakai and Kisumu also found that circumcision

decreases HPV incidence [RR = 0.67 (0.51-0.89) and HR = 0.61 (0.50-0.70)

(20)

The RCT in Rakai also examined the effect of male circumcision on HPV

infection in female partners. In HIV-negative men, circumcision was found to decrease the prevalence and incidence of HPV in their female partners [PRR =0.72 (0.62 -0.85)] (39). However, in HIV-positive men, circumcision was found to have no effect on transmission of HPV to female partners (40).

The Role of HPV Viral Load in the Natural History of HPV infections

Determinants of HPV Viral load in males

The studies done on HPV viral load in men so far have been cross-sectional and focused on the distribution and factors associated with viral load. Flores et al found that the penile shaft specimens had the highest viral load values than any another penile site, suggesting that the penile shaft is probably the preferred site for viral replication. Xi et al found that higher viral load values for types 16 and 18 were associated with current smoking status at baseline (41). Other factors such as number of lifetime partners, young age, STI infections, and condom use are known risk factors for HPV infection in men, but their association with HPV viral load remains unclear (24, 27, 28).

HPV Viral Load and HPV Transmission from men to women

The current understanding of the role of HPV viral load in transmission of HPV is based on a study of 238 heterosexual couples examining concordance and viral load (2). This study longitudinally examined HPV type concordance in females based on the viral load of their male partners. Men with detectable viral load levels (>250

(21)

stratified by HPV type, the association between viral load in men and concordance was observed for HPV 16 but not for HPV 18 and 31. This study showed that men with higher viral loads, particularly for HPV type 16, are more likely to transmit infection to their partners than men with lower viral loads.

HPV Viral Load and HPV Persistence or Clearance in women

The current published studies on HPV viral load and persistence or clearance of infection have been done in females; to my knowledge, our study would be the first to examine this association in men. Several studies in females have found that higher HPV viral loads for high risk HPV types are associated with persistence of infection (33-36). In these studies, viral load in women was been examined mainly in 2 ways: (binary as less than or greater than 100pg/ml) or as a continuous variable. In the studies that examined viral load dichotomously, the effect of high baseline viral loads on persistence was found to range from OR of 1.3 – 5.7 (13, 42, 43). One study that examined the effect of tertiles of viral load on persistence found that the rate of clearance was higher for patients with lower viral load tertiles than higher ones [HR =2.8 (1.0-8.1)](12). A study examining HPV viral load continuously found that women with ≥ 107 HPV 16 copies were infected for longer periods of time than women with lower viral load (11).

More recently, some studies have focused on the dynamics of viral load rather than single time point measurements. Marks et al examined viral load in 50 women and found that whereas the baseline viral load was not associated with clearance, a >2 log decline in viral load over 6 months (2 study visits) was strongly associated with

(22)

Even though the studies on viral load and persistence or clearance in women differ with regards to HPV types examined, viral load tests used, categorization of viral load and definitions of persistence or clearance, they suggest that lower viral loads are associated with clearance of infections in women, particularly for HPV 16.

HPV Viral load and development of CIN in women

Several studies that have examined the association between viral load and CIN have found greater viral load levels to be associated with greater severity of disease. Studies that examined types 16,18, and 31 found a strong association between high viral load (33rd percentile) was found to be strongly associated with ≥ CIN 2 (17, 45). A dose response-relationship has also been found for this association; the RR for

association between viral load and CIN3 was found to be 1.9 and 4.5 for 1-10 copies/ cell and >1000copies/cell of viral load respectively (14). A few studies have examined the association between the dynamics of viral load and CIN, these studies found that women who developed CIN2/3 had consistently high or increasing viral load values, whereas those who didn’t develop disease had decreasing viral load values over time (46, 47).

Biological Plausibility for Proposed Hypotheses

Aim 1: Male Circumcision reduces acquisition and enhances clearance of high viral load HPV infections

(23)

of the penis (39). Male circumcision is believed to enhance clearance by eliminating the moist subpreputial cavity that promotes HPV virus survival and subsequently reduces HPV clearance (39).

Aim 2: High HPV viral load is associated with lower HPV clearance

(24)
(25)

CHAPTER 3 METHODS

These analyses use data from two studies: an RCT of the effectiveness of male circumcision in reducing HIV incidence, and an ancillary HPV study conducted in men in Kisumu, Kenya from 2002-2007(5, 18). Penile exfoliated cells were collected from the men every 6 months for 2 years. Presence of human DNA was evaluated by beta-globin specific PCR, followed by agrose gel electrophoresis (19). The GP5+/6+ PCR assay was used to identify a wide array of HPV DNA types (19, 20). HPV viral load testing was conducted using a LightCyler real-time PCR assay; viral load was classified as high or low (>250 versus 250 copies/scrape) (21). The 250 copies/scrape cut-point was chosen because high HPV viral load (>250 copies/scrape) in men has been associated with increased HPV type concordance in female partners (2).

RCT of Male Circumcision for HIV Prevention: Bailey et al, 2007

(26)

high prothrombin time index, absolute indication for circumcision and other medical conditions contraindicating surgery.

Screening Visit

Participants were referred to the RCT by private and public clinics, and peer outreach workers recruited participants from local youth organizations. At the screening visits:

 Potential participants were interviewed to determine age, residence and sexual activity history

 Blood test for hemoglobin (to determine if ≥ 90g/l) were done

 HIV testing and counseling was provided

 A physical examination was also done to determine circumcision status. All eligible individuals were invited to participate in the study and were given a consent form in their preferred language (English, Dholuo or Kiswahili) to take with them and read in detail.

Baseline Visit

Trained study personnel reviewed the consent form and consenting adults were enrolled in the study. At the baseline or randomization visit:

 An interview was conducted to obtain medical history, socio-demographic and health information

(27)

Nairobi. HSV-2 antibodies were tested for in sera using an enzyme-linked immunosorbent assay (ELISA) (Kalon).

 Urine samples were obtained and tested for N. gonorrhea, T. vaginalis and C. trachomatis by PCR (Roche Diagnostics)

 Participants were assigned to the circumcision (intervention) or delayed circumcision (control) groups. A total of 2784 participants were enrolled in the RCT, 1391

participants were assigned to the circumcised group and 1393 were assigned to the intervention group. Circumcision was performed on the same day or within a few days of randomization; participants were checked for complications 3, 8 and 30 days after circumcision.

 Counseling for HIV and STI prevention was done

Follow-up visits

Participants were followed every 6 months at approximately 6, 12, 18 and 24 months after the baseline visit. At the follow-up visits:

 An interview was conducted to obtain socio-demographic and health information

 Counseling for HIV and STI prevention was done

 STI testing (as done at baseline) was done and treatment was provided

Design of HPV Nested Study

(28)

consent for the collection of penile exfoliated cell specimens at each visit. Of the 2784 participants enrolled in the parent RCT, 2299 (83%) consented to participate and were enrolled in the ancillary HPV study.

Collection of penile exfoliated cell specimens

(29)

Each swab was placed in a sterile 15 ml centrifuge tubes with 2-mL of 0.01 mol/L Tris-HCl 7.4 pH buffer with the participant’s ID number, the visit number and type of specimen (glans or shaft) (18).

Processing and shipping of collected specimens

Each sample tube (2 per participant, glans and shaft) was placed in a previously labeled centrifuge tubes processed and centrifuged at high speed for 10 minutes in the clinic laboratory on the same day as sample was collected. After centrifuging, the supernatant was discarded using a Pasteur pipette, and the pellet was resuspended in the same volume of 0.01mol/L of Tris-HCl buffer and vortexed. Diluted cell pellets were frozen and stored at -75°C. All samples were transported via Fedex in a dry shipper (with permission from the Kenyan Ministry of Health) from Kisumu to the Department of Pathology at the VU Medical Center in Amsterdam, the Netherlands for HPV DNA testing (5, 18).

HPV DNA testing

(30)

types: HPV6, 11, 16, 18, 26, 30, 31, 32, 33, 34, 35, 39, 40, 42, 43, 44, 45, 51, 52, 53, 54, 55, 56,57, 58, 59, 61, 64, 66, 67, 68, 69, 70, 71 (equivalent to CP8061), 72, 73, 81 (equivalent to CP8304), 82 (IS39 and MM4subtypes), 83 (equivalent to MM7), 84

(equivalent to MM8), cand85, 86, cand89 (equivalent to CP6108) and JC9710(19). HPV genotyping was subsequently performed on GP5+/6+ PCR positive PCR products by reverse line blot (RLB) hybridization (19, 20). For all analyses, HPV16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68 were considered high-risk types. HPV types detected by EIA but not by RLB genotyping were designated as HPVX, indicating a type, sub-type or variant not detectable with probes used for reverse line blot

hybridization. Low-risk types included all other HPV types.

HPV16/18 DNA viral load testing

(31)

Validity of viral load measures

The LightCycler real time PCR assay used have previously been compared to another real time PCR method (Taq Man assay) and the GP5+/6+ enzyme

immunoassay (EIA) to assess validity of measurements obtained (21). The results from the validity study showed that the LightCycler method viral load results were strongly correlated to those from the TaqMan assay (spearman ρ = 0.87), but less correlated to those from the EIA assay (spearman ρ = 0.77). These results showed that viral load results obtained with LightCylcer are comparable to those from other real time PCR methods; these real time PCR methods are more accurate in measuring viral load than EIA methods.

Results from the duplicate runs done per sample were obtained for baseline samples and assessed for agreement. The spearman coefficients for HPV16 and HPV18 were found to be 0.97 and 0.96 respectively (unpublished data). These results were limited in assessing for variability since they were done in the same run and do not account for technical variability, but they show very high agreement in viral load

measures produced by the LightCycler.

Statistical Analyses

Aim 1: Effect of male circumcision on incidence and clearance of high and low viral load HPV16 and HPV18 infections in Kenyan men

All analyses were performed for each HPV type (HPV16 or HPV18) and

(32)

Men who were randomized to the circumcision arm but were not circumcised (N=23) and those randomized to the control arm but were circumcised (N=14) were included in the analysis in the group to which they were randomly assigned. Incident infections were assumed to occur at the mid-point between the last HPV-negative and first subsequent HPV-positive result. Men who tested HPV-negative at each visit were censored at their last observed visit. Only the first incident infections for each HPV type (HPV16 or HPV18) at each anatomical site (glans or shaft) were included in incidence analyses. The Kaplan-Meier method was used to estimate the 24-month cumulative risk of HPV16 and HPV18 in the glans or shaft for the circumcision versus control arms. Hazard ratios (HRs) comparing the rate of acquisition of high viral load infections in circumcision versus control arms were estimated by fitting Cox proportional hazards models.

An additional analysis was conducted among only participants with incident HPV16 or HPV18 infections to examine associations between circumcision and the prevalence of high versus low HPV viral load at time of HPV detection. Potential confounders of the association between circumcision and HPV viral load were identified from literature, and analyzed using a directed acyclic graph (DAG) (50). Confounders identified included: HIV infection, condom use, number of sex partners, bathing frequency, infection with multiple HPV types and infection with other STIs (HSV-2, Trichomonas vaginalis, Chlamydia trachomatis and Neisseria gonorrhoeae). Log binomial models were used to derive

(33)

Clearance of prevalent infections detected at enrollment was assumed to occur at the mid-point between the last HPV-positive result and the first subsequent HPV-negative result. Men were censored at their last observed visit if they tested positive at each

consecutive visit for the same HPV type at the same anatomical site. The Kaplan-Meier method was used to estimate the cumulative risk of clearance for high and low viral load prevalent HPV infections. Risk estimates for clearance of prevalent infections in the circumcision versus control arms were compared with a Z test at 6 and 12 months after baseline. HPV persistence was defined as the detection of the same HPV type at the same anatomical site at 6 and 12 months after baseline in participants who were positive for HPV16 or HPV18 at baseline. For all the analyses of incidence and clearance, HR estimates accounting for interval censoring were obtained by fitting models assuming Weibull distributions for incidence times; results obtained were similar to those from the Cox models (data not shown). Sensitivity analyses were performed by restricting analyses to samples that tested positive for the presence of human DNA with the beta-globin test.

Aim 2: Association between HPV Viral Load and HPV clearance in uncircumcised Kenyan men

The distribution of participant characteristics by HPV16 or HPV18 viral load (high or low) was examined for infections prevalent at baseline. For these analyses, HPV viral load results from the glans and shaft samples were combined; high viral load was defined as >250 copies/scrape in the glans or shaft, and low viral load was defined as  250

(34)

HPV clearance analyses were performed separately for incident and prevalent HPV16 or HPV18 infections. We also examined HPV clearance separately for glans and shaft infections since HPV clearance has been shown to vary by anatomical site (8) . For each participant, only the first detected HPV16 or HPV18 infection (incident or prevalent) at each anatomical site (glans or shaft) was included in clearance analyses. Time-to-clearance since HPV detection was calculated starting at the visit date of first HPV detection; for prevalent infections the start date was baseline, and for incident infections it was the visit date of first HPV detection for each HPV type at each anatomical site. Clearance was assumed to occur at the mid-point between the last HPV-positive result and the first subsequent HPV-negative result. Men were censored at their last observed visit if they remained HPV positive for the same HPV type at the same anatomical site for each observed consecutive visit.

HPV viral load measurements at first detection of HPV16 or HPV18 prevalent or incident infections were used to classify the infections as high (>250 copies/scrape) or low ( 250 copies/scrape) viral load infections. For all the clearance analyses, HPV viral load was analyzed as a time-fixed exposure; few HPV16 or HPV18 infections (8.4% overall) were observed to change viral load groups during follow-up. The actual circumcision status of eligible men at each visit was reviewed to verify that they were uncircumcised during HPV detection and clearance; 12 men assigned to the control arm were circumcised later in the study, however, in each case circumcision was performed after clearance of HPV infections so these men were included in clearance analyses.

(35)
(36)
(37)

CHAPTER 4

THE ASSOCIATION BETWEEN MALE CIRCUMCISION AND PENILE HPV VIRAL LOAD IN AN RCT IN KISUMU, KENYA

Overview

(38)

[0.55(0.39-0.68)] for HPV16 and HPV18 [0.17(0.05-0.34) and 0.50(0.25-0.71), respectively]. Results from shaft samples were similar to those from glans samples. Circumcision reduces

acquisition and enhances clearance of high viral load HPV infections in the glans, and thus could potentially reduce HPV transmission to women.

Introduction

Persistent infections with human papillomavirus (HPV) high-risk types 16 and 18 are causes of approximately 70% of cervical cancer cases in women (1). Women with higher HPV16 or HPV18 viral loads are more likely to have persistent infections and to progress to high grade cervical intraepithelial neoplasia (CIN) than those with lower viral loads (11-17). Men with higher HPV16 or HPV18 viral loads have a higher prevalence of flat penile lesions and greater HPV type concordance with their female partners than those with lower viral loads (2, 52, 53). As a result, higher HPV viral load in men is suggested to be associated with increased HPV transmission to their female partners (2).

(39)

transmission to women, but the association between circumcision and HPV viral load remains unclear (2).

To our knowledge, only the RCT in Rakai, Uganda, has examined the effect of male circumcision on penile HPV viral load (22). This study, however, did not have longitudinal data on HPV viral load and was therefore unable to examine the association between circumcision and incident high versus low viral load HPV infections. The association between circumcision and the clearance of high and low viral load HPV infections present at baseline (prevalent infections) was also not explored.

We present the results of an RCT examining the associations between male

circumcision and the incidence and clearance of high and low viral load infections for HPV types16 and 18 in men in Kisumu, Kenya. We have previously shown that male

circumcision reduces HPV incidence and increases HPV clearance in this RCT (8); here we expand on those findings to examine the association between male circumcision and HPV viral load.

Methods

Study population and enrollment

(40)

assigned to either the immediate circumcision arm or to the control arm and asked to remain uncircumcised until the end of their 24 months of study participation. This study protocol was approved by Institutional Review Boards of the Universities of Illinois at Chicago, Manitoba, Nairobi and North Carolina; RTI International and VU University Medical Center.

This analysis includes men from the RCT who consented to the collection of penile exfoliated cells and their shipment overseas for HPV DNA and HPV viral load testing. Of the 2784 men enrolled in the RCT, 2299 (83%) consented to provide penile swab samples and had baseline HPV data; 1159 men were in the circumcision and 1140 in the control arm (Figure 4.I). Eligibility for inclusion in incidence and clearance analyses was

determined for each HPV type (HPV16 or 18) and anatomical site (glans or shaft)

(41)

Follow-up and specimen collection

Participant follow-up visits were scheduled every 6 months for a period of 24 months. A standardized questionnaire on sociodemographic characteristics and sexual behavior was administered by a trained male interviewer at each visit (5). Penile exfoliated cell specimens were collected by a trained physician or clinical officer at each study visit (18). Samples were collected from two anatomical sites: (i) penile shaft and external foreskin (shaft specimen); and (ii) glans, coronal sulcus, and the internal tissue of the foreskin (glans specimen). Sampling of the foreskin was conducted only in uncircumcised men. Two pre-wetted type 3 Dacron swabs were used to collect penile exfoliated cells from the two separate anatomical sites. Each swab was placed in a sterile 15 ml centrifuge tubes with 2-mL of 0.01 mol/L Tris-HCl 7.4 pH buffer with the participant’s ID number, the visit number, and anatomical site of specimen (glans or shaft). Each sample was

centrifuged at high speed for 10 minutes in the clinic laboratory on the same day it was collected. Excess Tris-HCl buffer was discarded using a Pasteur pipette, and the pellet was resuspended in the same volume of 0.01mol/L of Tris-HCl buffer and vortexed. Diluted cell pellets were frozen and stored at -75°C. All samples were transported via Fedex in a liquid nitrogen dry shipper from Kisumu to the Department of Pathology at VU Medical Center in Amsterdam, Netherlands, for HPV DNA testing.

HPV DNA, HIV and STI Testing

(42)

assessed by GP5+/6+ PCR followed by hybridization of PCR products using an enzyme immunoassay readout with two HPV oligoprobe cocktail probes that detect 44 HPV types. Subsequent HPV genotyping was performed by reverse line blot (RLB) hybridization of PCR products (19, 20).

HIV testing was performed at each visit using two HIV antibody rapid tests (Determine, Abbott Diagnostic Division, Hoofddorp, Netherlands; and Unigold, Trinity Biotech, Wicklow, Ireland), and confirmed by double ELISA (Adaltis, Montreal, Canada; Trinity Biotech, Wicklow, Ireland) at the University of Nairobi. Urine samples at each visit were tested for Trichomonas vaginalis, Neisseria gonorrhoeae and Chlamydia trachomatis

infections by PCR-based methods (Roche Diagnostics), and serum was tested for Herpes simplex virus (HSV)-2 antibodies by a type-specific enzyme-linked immunosorbent assay (ELISA) (Kalon).

HPV Viral Load Testing

(43)

chosen because high HPV viral load (>250 copies/scrape) in men has been associated with increased HPV type concordance in female partners (2).

Statistical Analyses

All analyses were performed for each HPV type (HPV16 or HPV18) and

anatomical site (glans or shaft) separately. Intent-to-treat analyses were performed to examine the incidence of high viral load infections in circumcision versus control arms. Men who were randomized to the circumcision arm but were not circumcised (N=23) and those randomized to the control arm but were circumcised (N=14) were included in the analysis in the group to which they were randomly assigned. Incident infections were assumed to occur at the mid-point between the last HPV-negative and first subsequent HPV-positive result. Men who tested HPV-negative at each visit were censored at their last observed visit. Only the first incident infections for each HPV type (HPV16 or HPV18) at each anatomical site (glans or shaft) were included in incidence analyses. The Kaplan-Meier method was used to estimate the 24-month cumulative risk of HPV16 and HPV18 in the glans or shaft for the circumcision versus control arms. Hazard ratios (HRs) comparing the rate of acquisition of high viral load infections in circumcision versus control arms were estimated by fitting Cox proportional hazards models.

(44)

condom use, number of sex partners, bathing frequency, infection withmultiple HPV types and infection with other STIs (HSV-2, Trichomonas vaginalis, Chlamydia trachomatis and

Neisseria gonorrhoeae). Log binomial models were used to derive prevalence ratios (PRs) comparing the prevalence of high versus low viral load at the time of HPV detection for incident infections observed in the circumcision versus control arms.

Clearance of prevalent infections detected at enrollment was assumed to occur at the mid-point between the last HPV-positive result and the first subsequent HPV-negative result. Men were censored at their last observed visit if they tested positive at each

consecutive visit for the same HPV type at the same anatomical site. The Kaplan-Meier method was used to estimate the cumulative risk of clearance for high and low viral load prevalent HPV infections. Risk estimates for clearance of prevalent infections in the circumcision versus control arms were compared with a Z test at 6 and 12 months after baseline. HPV persistence was defined as the detection of the same HPV type at the same anatomical site at 6 and 12 months after baseline in participants who were positive for HPV16 or HPV18 at baseline. For all the analyses of incidence and clearance, HR estimates accounting for interval censoring were obtained by fitting models assuming Weibull distributions for incidence times; results obtained were similar to those from the Cox models (data not shown). Sensitivity analyses were performed by restricting analyses to samples that tested positive for the presence of human DNA with the beta-globin test.

Results

(45)

reported having more than 2 sex partners and 22% reported always using a condom in the previous 6 months. The prevalence of other STIs at baseline was low (2% for Trichomonas vaginalis, 2% for Neisseria gonorrhoeae and 5% for Chlamydia trachomatis), with the exception of HSV-2 (27%). There were no appreciable differences in sociodemographic, sexual history, or STI laboratory results at baseline between eligible participants

randomized to the circumcision and control arms (Table 4.1).

Incidence of high viral load HPV infections

The 24-month cumulative risk of high viral load infections in the glans was lower in the circumcision than control arm for HPV 16 (2.7% versus 9.5%) and HPV18 (2.2% versus 6.8%, respectively) (Figure 4.2). In the shaft, the estimated 24-month cumulative risk of high viral load infections was comparable in the circumcision and control arms for HPV16 (2.0% versus 2.6%) and HPV18 (2.0% versus 2.6%, respectively).

Of incident HPV infections observed, a total of 151 HPV16 (39.7%) and 135 HPV18 (69.2%) high viral load infections occurred in either the glans or shaft (Table 4.2). The rate of acquisition of high viral load infections in the glans was lower in the circumcision than control arm for both HPV16 [HR=0.32 (0.20-0.49)] and HPV18 [HR= 0.34 (0.21- 0.54)] (Table I). Similar trends were observed in the shaft for both HPV types [HPV16 HR= 0.79 (0.42 -1.45) and HPV18 HR= 0.87 (0.48-1.54)]. Sensitivity analysis by restricting to beta-globin-positive samples yielded similar results (Table 4.2).

(46)

for HPV18]. In the shaft, most HPV16 incident infections had low viral loads [74.3% (N=52) in circumcision arm; 72.4 %( N=62) in control arm] whereas most HPV18 incident infections had high viral loads [77.8% (N=21) in circumcised arm; 65.0% (N=26) in the control arm]. Few men had incident infections in both the glans and shaft (19% for HPV16 and 25% for HPV18), and in all these cases the viral load category (high or low viral load) was the same in both anatomical sites.

Among men with incident HPV infections in the glans, the prevalence of high versus low viral load at the time of HPV detection was lower in the circumcision than control arm for HPV16 [PR=0.70(0.50-0.99)] and HPV18 [PR=0.59(0.43-0.82)]. However, this

association was not observed in the shaft for HPV 16 [PR = 0.93 (0.55-1.57)] or HPV 18 [PR = 1.20 (0.88-1.62)] (Table 4.3). PRs obtained after adjustment for HIV infection, condom use, number of sex partners, bathing frequency, infection with STIs and multiple HPV type infection were similar to the reported unadjusted HRs (results not shown).

Sensitivity analyses restricted to beta-globin-positive samples showed similar results as the main analysis (Table 4.3).

Clearance of high and low viral load infections prevalent at baseline

For both HPV types, HPV prevalence at baseline was higher in the glans (HPV16 =6.6%; HPV18 =3.1%) than in the shaft (HPV16 =2.8%; HPV18 =1.1%). Prevalent

infections in the glans at baseline were more likely to have high HPV viral loads [HPV16= 54.9% (79); HPV18= 66.7% (N=40) than prevalent infections at baseline in the shaft

(47)

prevalent HPV16 and HPV18 infections at baseline were assigned to the circumcision and control arms (Figure 4. I).

Nearly all prevalent infections at baseline cleared during follow-up, with the

exception of 5 (6.6%) HPV16 glans infections in the control arm (4 high viral load infections censored at 18 months and 1 low viral load infection censored at 12 months after baseline). For both HPV types, most prevalent infections in the glans or shaft cleared within 6 months (Figures 4.3and 4.4). At both 6 and 12 months after baseline, the highest risk of

persistence for HPV16 and HPV18 infections was among men in the control arm with high HPV viral load. The estimated risk of persistence to 6 months for prevalent infections in the glans was lower in the circumcision arm [0.20(0.09-0.34)] than control arm [0.55(0.39-0.68)] for HPV 16 and HPV18 [0.17(0.05-0.34) and 0.50(0.25-0.71), respectively] (Table4.4). For prevalent HPV16 infections in the glans, the estimated risk of persistence to 6 months for high viral load infections in the circumcision arm [0.20 (0.09 – 0.34)] was similar to that of low viral load infections in the control arm [0.13 (0.04 -0.26)] (Table 4.4). A similar trend was observed for prevalent HPV16 infections in the shaft and prevalent HPV18 infections in the glans and shaft.

(48)

(0.70-5.15)for HPV18]. Sensitivity analyses of clearance restricted to prevalent infections that were beta-globin-positive found similar results to the main analysis (data not shown).

Discussion

The two-year cumulative incidence of high viral load HPV16 and HPV18 infections in the glans was lower in the circumcision arm than control arm in young men from Kisumu, Kenya. Circumcision reduced the rate of acquisition of high viral load HPV16 and HPV18 infections in the glans over 24 months. Among men with incident HPV16 or HPV 18

infections, circumcised men were more likely to have low versus high viral load infections in the glans as compared to uncircumcised men. We also found that for prevalent high viral load infections detected at baseline, the risk of HPV persistence to 6 months was lower in the circumcision arm than control arm for both HPV types 16 and 18 in the glans. Similar, but less precise results were observed for the estimated effect of circumcision on the incidence of high viral load HPV16 and HPV18 infections in the shaft samples.

To our knowledge, only one study, in Rakai, Uganda, has examined the effect of circumcision on HPV viral load in men (22). Also an RCT of male circumcision, the study found the combined HPV viral load for HPV types 16, 18, 31, 33, 35 and 52 in the glans was lower in the circumcision arm than control arm at 24 months after randomization. Their estimated protective effect of circumcision on prevalence of high versus low HPV viral load infections in the glans [PR =0.54 (0.21-1.42)] was similar to the PR estimates we obtained for HPV16 [0.70 (0.50 -0.99)] and HPV18 [0.59 (0.43 -0.82)] in the glans.

(49)

viral load infections in the shaft were similar, but less precise than estimated effects found in the glans. This may be due to fewer HPV infections and a lower proportion of high viral load infections detected in the shaft as compared with the glans. Stronger effects of

circumcision on HPV infection in the glans than more distal anatomical sites to the foreskin have been documented in other studies (37).

The biological mechanism by which circumcision reduces HPV viral load is unclear, but it is believed that the keratinized skin surface and scar tissue in circumcised men reduces the entry and subsequent HPV viral replication in the basal epidermal cells of the penis (39). In uncircumcised men, it has been hypothesized that the moist subpreputial cavity promotes HPV virus survival and subsequently reduces HPV clearance (39, 54). Male circumcision has been previously shown to reduce the prevalence of flat penile lesions; our findings are consistent with the current understanding that circumcision

reduces the prevalence of flat penile lesions by reducing HPV viral load in men (2, 52, 53). Furthermore, the effect of circumcision in reducing HPV viral load may explain why fewer HPV infections and lower HPV viral loads are observed in female partners of circumcised than uncircumcised men (40, 54). It is possible that circumcision reduces the transmission of HPV infections to women by reducing incidence and enhancing clearance of high viral load HPV infections in their male partners.

(50)

copies/scrape units have been previously used to show that high HPV viral load (>250 copies /cell) in men is associated with greater HPV type concordance in their female partners (2). We also observed fairly low beta-globin positivity in our samples (61.5% in glans and 43.5% in shaft samples). A possible explanation is that penile cells, particularly in shaft samples, are more keratinized and anucleated than those in the cervix, and therefore may contain less human DNA (24). Lower beta-globin positivity in penile swab samples has also been documented in a similar study (7). Due to our low beta-globin positivity, it is possible that HPV viral load measurements in this study may be underestimated. However, all the incidence and clearance sensitivity analyses restricted to beta-globin positive

samples found similar results as the main analyses.

This study adds to the current understanding of HPV viral load in men in general, and the natural history of HPV infections in circumcised and uncircumcised men. Like previous studies examining HPV viral load in men, the interpretation of our results is limited by the inability to distinguish between integrated and episomal DNA (48). Integration is followed by a decrease in HPV viral load, but the level of integrated HPV is believed to increase with progression of HPV infections (48). Future research distinguishing between episomal and integrated DNA is needed to further understand the effect of circumcision on HPV viral load.

(51)

independently of circumcision, although this needs to be confirmed in a larger study taking into account other factors associated with HPV clearance including treatment of other STIs.

(52)

Tables and Figures

Figure 4.1: Study flowchart of HPV samples included in analyses of the effect of male circumcision on penile HPV viral load

HPV Types HPV 16 HPV 18

Anatomic Sites Glans Shaft Glans Shaft

Other Exclusion Criteria :

Missing HPV viral load data 19 11 2 1 HPV (+) at baseline

Total samples in

Clearance Analyses 68 39 37 12

HPV (-) at baseline

Total Samples in Incidence

Analyses 1009 1046 1057 1083

HPV 16 HPV 18

Glans Shaft Glans Shaft

7 6 10 2

76 28 23 12

1014 1063 1064 1083

1159 men in Circumcision Arm

2784 men randomized in circumcision trial (Bailey et al, 2007)

1140 men in Control Arm 2299 men consented with baseline HPV data

485 men did not consent to HPV ancillary study or had no baseline HPV data

63 no follow-up HPV data

43 no follow-up HPV data

1096 men with HPV follow-up data

1097 men with HPV follow-up data

(53)

Figure 4.2: Estimated Cumulative Incidence of High Viral Load infections over 24 months in Kenyan Men

HPV16 GLANS HPV16 SHAFT

Number at Risk

Circumcision Arm

Control Arm 1009 1014 987 956 947 891 812 888 663 755 1046 1063 1028 1037 982 981 920 903 784 760

HPV18 GLANS HPV18 SHAFT

Number at Risk Circumcision Arm Control Arm

1057 1045 1004 945 815

1064 1039 991 903 761 1083 1083 1071 1041 1068 1028 985 956 852 812

Log rank P < 0.001 Log rank P = 0.46

Log rank P < 0.001 Log rank P = 0.53

(54)

Figure 4.3: Estimated Clearance of Prevalent High Viral Load HPV Infections by circumcision status in Kenyan Men

HPV16 GLANS HPV16 SHAFT

Number of Prevalent HPV Infections

Circumcision Arm

Control Arm 35 44 7 24 2 8 4 0 - - 5 3 0 2 0 1 0 0 - -

HPV18 GLANS HPV18 SHAFT

Number of Prevalent HPV Infections

Circumcision Arm

Control Arm 24 16 4 8 1 4 1 0 0 0 4 6 0 1 0 0 - - - -

Log rank P = 0.006

Log rank P = 0.125

Log rank P = 0.036

Log rank P = 0.055

(55)

Figure 4.4: Estimated Clearance of Low Viral Load Prevalent HPV Infections by circumcision status in Kenyan Men

HPV16 GLANS HPV16 SHAFT

Number of Prevalent HPV Infections

Circumcision Arm

Control Arm

33 0 0 -

32 4 1 - 34 25 5 2 0 1 0 0

HPV18 GLANS HPV18 SHAFT

Number of Prevalent HPV Infections

Circumcision

Arm 13 7 1 1 0 0 - - 8 6 0 0 0 0 - -

Log rank P = 0.387 Log rank P = 0.440

Log rank P = 0.223

Log rank P = 0.033

(56)

Table 4.1: Baseline Characteristics by Intervention Treatment Arm

Circumcision Arm (N= 1096*) Control Arm (N= 1097*) Participant Characteristics †

Age (years) 20(19-22; 18-28; 1096) 20(19-22; 17-24; 1097)

Education level

Less than secondary

Any secondary or above 372 (33.9) 724 (66.1) 391 (35.6) 706 (64.4)

Employment status

Employed and receiving a salary Self-employed

Unemployed

94 (8.6) 303 (27.7) 699 (63.8)

98 (8.9) 294 (26.8) 705 (64.3)

Marital status

Not married (no live-in partner) Not married (with live-in partner) Married (not living with wife) Married (living with wife)

1024 (93.8) 7 (0.6) 5 (0.5) 56 (5.1) 1018 (93.2) 7 (0.6) 15 (1.4) 52 (4.8)

Number of partners over lifetime 4 (3-7; 1-120; 1004) 4 (3-7; 1-86; 1016)

Number of sexual partners in previous 6 months 0 1 2+ 137 (12.5) 472 (43.2) 484 (44.3) 136 (12.4) 472 (44.1) 475 (43.5)

Used a condom with intercourse in previous 6 months Always Inconsistent Never 210 (21.9) 511 (53.3) 238 (24.8) 208 (21.7) 500 (52.1) 251 (26.2)

Herpes simplex virus 2 Positive

Negative 287 (27.3) 764 (72.7) 278 (26.5) 771 (73.5)

Trichomonas vaginalis

Positive

Negative 1064 (98.1) 21 (1.9) 1059 (97.8) 24 (2.2)

Neisseria gonorrhoeae

Positive

Negative 1053 (97.2) 30 (2.8) 1066 (98.4) 17 (1.6)

Chlamydia trachomatis

Positive Negative

57 (5.3) 1025 (94.7)

42 (3.9) 1041 (96.1)

Bathing frequency Less than daily Daily

23 (2.1) 1063 (97.9)

22 (2.0) 1062 (98.0)

(57)

*This table examines baseline characteristics in men with HPV follow-up data randomized to the circumcision and control arms

Data are median (IQR; range; n) for continuous data, or n (%) for categorical data. Numbers and percentages may not sum up to totals

due to missing values and rounding.

(58)

Table 4.2: Effect of circumcision on incident high viral load HPV16 and HPV18 infections over 24 months in Kenyan Men

HPV

Type Site

Treatment Arm

All Samples Beta-globin Positive Samples

Incident High VL Infections* Perso n Years

HR (95% CI) Incident High

VL Infections*

Person

Years HR (95% CI)

HPV 16

Glans Circumcision 26 1386.7 0.32 (0.20-0.49) 19 875.7

0.31 (0.18 – 0.50)

Control 83 1366.1 1 72 976.0 1

Shaft Circumcision 18 1443.8 0.79 (0.42 -1.45) 9 689.4 0.76 (0.31 -1.76)

Control 24 1481.1 1 13 703.4 1

HPV 18

Glans Circumcision 22 1473.0

0.34 (0.21 –

0.54) 15 902.6 0.40 (0.21 -0.70)

Control 66 1486.8 1 45 1055.4 1

Shaft Circumcision 21 1521.8

0.87 (0.48 –

1.54) 11 725.1 0.67 (0.31 -1.42)

Control 26 1539.9 1 17 719.7 1

* Incidence was defined as the first type-specific HPV-positive result in men who were negative for that HPV type at the same anatomical site at baseline. High viral load (VL) was defined as >250 copies/scrape for the given HPV type.

† Hazard Ratio (HR) comparing the risk of incident high viral load infections in circumcision versus control arm.

(59)

Table 4.3: Association between circumcision and prevalence of high versus low viral load at detection of incident HPV infections

HPV

Type Site

Treatment Arm

All Samples Beta-globin Positive Samples

Incident HPV High VL/ Low

VL*

PR (95% CI) Incident HPV

High VL/ Low VL*

PR (95% CI)

HPV 16

Glans Circumcision 26 /43 0.70 (0.50 - 0.99) 19 /26

0.72 (0.53 - 0.98)

Control 83 / 72 1 72 /51 1

Shaft Circumcision 18 / 52 0.93 (0.55 - 1.57) 9 / 22

0.92 (0.45 – 1.86)

Control 24 / 62 1 13 /28 1

HPV 18

Glans Circumcision 22 /24 0.59 (0.43 – 0.82) 15/16 0.59 (0.40 -0.87)

Control 66 /16 1 45/10 1

Shaft Circumcision 21 / 6 1.20 (0.88 – 1.62) 11 / 3 1.25 (0.84 -1.86)

Control 26 /14 1 17/10 1

*Incidence was defined as the first type-specific HPV-positive result in men who were negative for that HPV type at the same anatomical site at baseline. High viral load (VL) was defined as >250 copies/scrape; low viral load was defined as  250 copies/scrape for the given HPV type.

Prevalence Ratio (PR) comparing the prevalence of high versus low viral load at HPV detection in men with incident HPV infections in the circumcision versus control arm.

(60)

Table 4.4: Clearance of baseline prevalent infections by circumcision status and HPV viral load in Kenyan Men

Site LoadViral Treatment Arm infections n/N* Cleared Person Years

HR for Clearance (95% CI)

Estimated Risk of HPV Persistence (95% CI) ††

6 months 12 months

HPV16 Infections

Glans

High Circumcision 35/ 35 14.4 1.46 (0.91 -2.32) 0.20 (0.09 -0.34)** 0.06 (0.01-0.17)

Control 40/44 23.8 1 0.55 (0.39 -0.68)** 0.18 (0.09-0.31)

Low Circumcision 33/33 9.2 1.31 (0.68 -1.89) 0 0

Control 31/32 9.3 1 0.13 (0.04 -0.26) 0.03 (0.00 -0.14)

Shaft

High Circumcision 5/5 1.3 4.34 (0.67 -84.9) 0 0

Control 3/3 2.2 1 0.67 (0.05 -0.94) 0.33 (0.01 -0.77)

Low Circumcision 34/34 12.1 0.82 (0.49 -1.42) 0.15 (0.05 -0.28) 0

Control 25/25 7.8 1 0.08 (0.01 -0.22) 0.04 (0.00 - 0.17)

HPV18 Infections

Glans

High Circumcision 24/24 9.1 2.02 (1.02- 4.08) 0.17 (0.05 -0.34)** 0.04 (0.00 -0.18)

Control 16/16 10.8 1 0.50 (0.35 -0.71)** 0.25 (0.08 -0.47)

Low Circumcision 13/13 3.9 1.79 (0.70 -5.15) 0.08 (0.00 -0.29) 0

Control 7/7 2.5 1 0.14 (0.01 -0.46) 0

Shaft

High Circumcision 4/4 1.0 4.33 (0.78 -33.3) 0 0

Control 6/6 2.3 1 0.15 (0.01 -0.52) 0

Low Circumcision 8/8 2.5

0.34 (0.10 -1.12) 0 0

Control 6/6 1.5 1 0 0

High viral load was defined as >250 copies/scrape; low viral load was defined as 250 copies/scrape for the given HPV type. *Clearance was defined as an HPV-negative result in participants who were positive for that given HPV type at baseline; n= number of cleared HPV infections; N=total number of HPV infections present at baseline.

Hazard Ratio (HR) comparing clearance of prevalent HPV infections at baseline in circumcision versus control arm over 24 months.

††Persistence was defined as an HPV-positive result in participants who were positive for that given HPV type at the same anatomical site in

the previous visit. Estimates for risks of persistence and duration since baseline visit were obtained from Kaplan Meier graphs. **P-value

<0.05 for z-score test comparing estimated risk of HPV persistence in circumcision versus control arm.

(61)

CHAPTER 5

ASSOCIATION BETWEEN PENILE HPV VIRAL LOAD AND HPV CLEARANCE IN UNCIRCUMCISED KENYAN MEN

Overview

(62)

explain the higher incidence of penile carcinomas observed in the glans than any other penile site. The lower HPV16 clearance in men with high viral load infections may also explain the reported increase in HPV16 transmission to their female partners in

comparison to men with low viral load.

Introduction

Persistent infections with human papillomavirus (HPV) high-risk types 16 and 18 are causes of approximately 70% of cervical cancer in women and approximately half of penile squamous cell carcinomas (SCC) in men, particularly basaloid and warty SCC types (1, 9, 10, 33). Women with higher HPV16 or HPV18 viral loads are more likely to have persistent infections and to progress to high grade cervical intraepithelial

neoplasia (CIN) than those with lower viral loads (11-17). The association between HPV viral load and HPV clearance has not yet been examined in penile HPV infections in men; higher HPV16 or HPV18 viral loads could possibly be associated with lower HPV clearance and increased progression to penile SCC.

Higher HPV16 or HPV18 viral loads in men are associated with an increased prevalence of flat penile lesions and greater HPV type concordance with female

partners in comparison to lower viral loads (2, 52, 53). As a result, higher HPV viral load in men is suggested to be associated with increased HPV transmission to their female partners in comparison to lower HPV viral load (2). It is believed that lower HPV

(63)

We present the results of a longitudinal study examining the association between viral load and the clearance of penile HPV infection in uncircumcised men in Kisumu, Kenya. We examine penile HPV viral load measured at first detection of HPV16 or HPV18 infections in the glans and shaft sites and the subsequent rate of HPV clearance over a study period of 24 months. The association between HPV viral load and HPV clearance is examined separately for incident infections and infections detected at baseline (prevalent infections). We have previously examined the effect of male

circumcision on HPV viral load, which included men from this study in the control group (55); here we examine the association between penile HPV viral load and the clearance of penile HPV infections among these uncircumcised men.

Methods

Study population and enrollment

Uncircumcised men were screened between February 4, 2002 and September 6, 2005 in Kisumu, Kenya to participate in an RCT of male circumcision (5). The main aim of the trial was to determine the effectiveness of male circumcision in reducing HIV incidence. Baseline study inclusion criteria included being uncircumcised, aged 18-24, HIV negative, sexually active within the past 12 months and having blood hemoglobin ≥ 90 g/l. Study participants were recruited from STI clinics, workplaces and community organizations. The study protocol was approved by Institutional Review Boards of the Universities of Illinois at Chicago, Manitoba, Nairobi and North Carolina; RTI International and VU University

(64)

This analysis includes men from the control arm of the RCT who consented to the collection of penile exfoliated cells and their shipment overseas for HPV DNA and HPV viral load testing. Of the 2784 uncircumcised men enrolled in the RCT, 2299 (83%) consented to provide penile swab samples and had baseline HPV data; 1159 men were randomized to immediate circumcision and 1140 to the control arm. Eligibility for inclusion in clearance analyses was determined for each HPV type (HPV16 or 18) and anatomical site (glans or shaft) separately. Participants assigned to the control arm who: i) had no HPV follow-up data; ii) were HPV16 or HPV18-negative throughout follow-up; iii) had their first HPV16 or HPV18-positive result at the last study visit; or iv) had missing HPV viral load data were excluded from clearance analyses for that specific HPV type and anatomical site (Figure 5.1.).

Follow-up and specimen collection

(65)

laboratory on the same day of collection. Excess Tris-HCl buffer was discarded using a Pasteur pipette, and the pellet was resuspended in the same volume of 0.01mol/L of Tris-HCl buffer and vortexed. Diluted cell pellets were frozen and stored at -75°C. All samples were transported via Fedex in a liquid nitrogen dry shipper from Kisumu to the Department of Pathology at VU Medical Center in Amsterdam, Netherlands, for HPV DNA testing.

HPV DNA, HIV and STI Testing

DNA was isolated from penile exfoliated cell samples using NucleoSpin 96 Tissue kit (Macherey-Nagel, Germany) and a Microlab Star robotic system (Hamilton, Germany) according to manufacturers’ instructions. Presence of human DNA was evaluated by beta-globin specific PCR, followed by agrose gel electrophoresis (19, 20).HPV positivity was assessed by GP5+/6+ PCR followed by hybridization of PCR products using an enzyme immunoassay readout with two HPV oligoprobe cocktail probes that detect 44 HPV types (high-risk: HPV16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68; low-risk: HPV6, 11, 26, 30, 32, 34, 40, 42, 43, 44, 53, 54, 55, 57, 61, 64, 67, 69, 70, 71, 72, 73, 81, 82, 83, 84,cand85, 86, cand89, JC9710). Subsequent HPV genotyping was performed by reverse line blot (RLB) hybridization of PCR products (19, 20).

(66)

collected for polymerase chain reaction (PCR) detection of C. trachomatis and N.

gonorrhea (Roche Diagnostics). Blood was collected for herpes simplex virus type 2 (HSV-2; enzyme-linked immunosorbent assay (ELISA) [Kalon]).

HPV Viral Load Testing

HPV DNA viral load testing was performed on penile exfoliated cell samples that contained HPV16 or HPV18 using a real time PCR assay and a LightCycler instrument (Roche, Mannheim, Germany) (21). DNA extraction and purification were conducted according manufacturer instructions using 100 μl of the remaining cell suspensions. All samples were run in duplicate in the same run and resulting viral load values were

averaged. Dilutions of cloned HPV DNA were used to determine the standard curve for the HPV target. To calculate the number of copies per sample, the amount of HPV DNA in one femtogram was multiplied by the dilution factor and divided by the weight of the cloned pHPV viral genome. HPV viral load was categorized as low (250 copies/scrape) or high (>250 copies/scrape) for each HPV16 and HPV18 infection at each anatomical site. The 250 copies/scrape cut-point for high versus low viral load was chosen because of its significance in previous HPV viral load studies; men with high viral load values (>250 copies/scrape) have been found to have a higher prevalence of flat penile lesions and increased HPV type concordance with female partners (2, 53).

Statistical Analyses

(67)

>250 copies/scrape in the glans or shaft, and low viral load was defined as  250

copies/scrape in both the glans and shaft. Baseline characteristics examined included age, education, employment, marital status, bathing frequency, number of sexual partners and condom use in the previous 6 months.

HPV clearance analyses were performed separately for incident and prevalent HPV16 or HPV18 infections. We also examined HPV clearance separately for glans and shaft infections since HPV clearance has been shown to vary by anatomical site (8) . For each participant, only the first detected HPV16 or HPV18 infection (incident or prevalent) at each anatomical site (glans or shaft) was included in clearance analyses. Time-to-clearance since HPV detection was calculated starting at the visit date of first HPV detection; for prevalent infections the start date was baseline, and for incident infections it was the visit date of first HPV detection for each HPV type at each anatomical site. Clearance was assumed to occur at the mid-point between the last HPV-positive result and the first subsequent HPV-negative result. Men were censored at their last observed visit if they remained HPV positive for the same HPV type at the same anatomical site for each observed consecutive visit.

Figure

Updating...

References

Updating...

Related subjects :